Reactions of Ar+ with Selected Volatile Organic Compounds. A

Jerry Hoff, John Herlinger, Tom Hickey, and Christopher M. Hadad*. Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue, Columbus,...
0 downloads 0 Views 128KB Size


J. Phys. Chem. A 2000, 104, 11318-11327

Reactions of Ar+ with Selected Volatile Organic Compounds. A Flowing Afterglow and Selected Ion Flow Tube Study Michael H. Cohen, Cynthia Barckholtz, Brian T. Frink, Joshua J. Bond, C. Michael Geise, Jerry Hoff, John Herlinger, Tom Hickey, and Christopher M. Hadad* Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue, Columbus, Ohio 43210-1173 ReceiVed: July 12, 2000; In Final Form: October 5, 2000

Temperature-dependent rate coefficients and branching ratios for the reactions of Ar+ with a variety of volatile organic hydrocarbons are reported. Reactions of N2+ and CO2+ were undertaken to calibrate a newly constructed variable-temperature flowing afterglow (VTFA) complemented with a selected ion flow tube (SIFT). In addition, the first determinations of rate coefficients for the reaction of Ar+ with several organic hydrocarbons (toluene, pyridine, furan, thiophene, cyclohexene, cyclooctene, cyclohexane, and tetrahydrofuran) have been measured between 298 and 423 K with a VT-SIFT. Analogous to the Ar+ + C6H6 reaction, these reactions proceed by nondissociative and dissociative charge transfer, and very little temperature dependence is observed for the rate coefficients. At 298 K and 0.5 Torr, the rate coefficients are similar [(1.4-1.8 ) × 10-9 cm3/s] for all of the Ar+ reactions with the organic hydrocarbons, except for pyridine (2.3 × 10-9 cm3/s). These values are in reasonable agreement with rate coefficients predicted from average dipole orientation (ADO) theory. The degree of fragmentation appears to be loosely correlated to the difference in ionization potential between Ar+ and that of the neutral compounds as well as the degree of unsaturation of the hydrocarbon.

I. Introduction The study of gas-phase ion-molecule reactions has made significant contributions to many areas of chemistry, including combustion, plasma, and ionospheric chemistry.1,2 One of the most commonly observed processes in gas-phase ion chemistry is charge transfer, in which a single electron is transferred from the neutral reagent to the reactant cation. Charge-transfer reactions have been studied by a variety of methods, including ion beam techniques3,4 and mass spectrometric techniques.5,6 The flowing afterglow (FA) technique, introduced by Ferguson, Fehsenfeld, and Schmeltekopf in 1963, is now an established method for studying the reaction kinetics of ionmolecule reactions.7 Initially, the FA contributed tremendously in the determination of quantitative rate coefficients for ionmolecule reactions of ionospheric and atmospheric interest. Although the FA method is very useful in the study of ionmolecule reactions under well-characterized conditions and temperatures, the technique is limited by the ability to generate the desired ion synthetically without the presence of any other ions. Furthermore, the ions are generated in a plasma that will contain positive ions, negative ions, and electrons, and these species may complicate the observed chemistry. As a result, there may be ambiguity in the identification of primary products from a specific reactant as well as the partial inability to distinguish between the desired reaction and other competitive processes. Smith and Adams realized that some of these problems could be resolved if the desired ion could be mass selected (“purified”) from an ion plasma generated in a remote source, and their efforts led to the development of the selected ion flow tube * Corresponding author. Fax: 614-292-1685. E-mail: [email protected]

(SIFT).8 The greatest challenge to its implementation was how to efficiently introduce the ions from the low-pressure, massselection region (typically 10-5 Torr) into the high-pressure reaction region (typically 0.5 Torr). This problem was first solved by the use of an injector flange that introduced the helium carrier gas through an inlet at the point of ion injection in such a manner that the pressure was reduced in the ion injection region.9 The NOAA laboratory10 demonstrated the benefit of a venturi inlet over the initial hole-pattern design, and the venturi design was exploited successfully by DePuy, Bierbaum, and co-workers.11 Since then, the widely recognized advantages of the SIFT have led to the adoption of the technique in several laboratories,11-18 and several detailed descriptions of the SIFT technique have been reported.9,19,20 Recently, Fishman and Grabowski have reported a flexible hole-pattern design and demonstrated its utility for some positive ion reactions.14 The most common design for current FA-SIFT instruments is a linear geometry,11 and as a result, the mass-selection chamber is on-axis with the ion source. One of the greatest limitations for quadrupole resolution is the presence of background gas in the quadrupole chamber. This effect can be minimized if the ion beam can be separated from the carrier gas introduced in the flowing afterglow source. A novel design21 for our FA-SIFT was chosen in an effort to reduce the effect that the neutral gas load would have on the sensitivity of the selection quadrupole. In the design presented here, an electrostatic quadrupole deflector (QD) is introduced into the ion path immediately before the mass-selection quadrupole. The QD thereby filters the ions away from the neutrals which have passed through the initial skimmer (nose cone orifice). The QD has long been used as a focusing device, but was first described in theory and applied in practice to merge oppositely charged ion beams by Zeman.22 Since then, the

10.1021/jp002489o CCC: $19.00 © 2000 American Chemical Society Published on Web 11/04/2000

Reactions of Ar+ with Volatile Organic Compounds

J. Phys. Chem. A, Vol. 104, No. 48, 2000 11319

Figure 1. General diagram of our variable-temperature flowing afterglow selected ion flow tube (VTFA-SIFT) with a neutral flow reactor.

quadrupole deflector has also been used to obtain the coaxial overlap of an ion and a laser beam,23,24 to merge oppositely charged clustered ion beams,25 and as a means to extract ions from different ion sources.26 The Zeman QD employed shims and cylindrical electrodes to approximate the hyperbolic field of the theoretical quadrupole deflector. The deflector quadrupole designed for our SIFT chamber is similar to that of Posey and co-workers;27 it was patterned after a simple design without shims originally described by Farley23 and further characterized by Mahaffy and Lai.26 Recently, the role of charge transfer and ionization in hydrocarbon combustion has been explored.28,29 Although many of the reactions that typically occur in combustion processes occur via neutral-neutral reactions, ion-molecule reactions are usually more facile as a result of the attractive potential of the ion. Recent results indicate that the inclusion of ionic mechanisms in combustion kinetic models increases the rate of combustion.30 To this end, Viggiano and co-workers have studied a variety of ion-molecule reactions with several aliphatic28,31 and a few aromatic hydrocarbons.29,32 With our current interest in combustion systems,33-39 we have also begun to study the chemistry of ions derived from atmospheric constituents with a variety of organic substrates of relevance to fuels. As a preliminary step, in this paper, we report the nondissociative and dissociative charge-transfer reactions of Ar+ with a series of organic hydrocarbons (benzene, toluene, pyridine, furan, thiophene, cyclohexene, cyclooctene, cyclohexane, and tetrahydrofuran). We present a brief summary of the design of the newly constructed VTFA-SIFT instrument that was used to study these reactions. Several well-established ion-molecule reactions involving atmospheric ions and VOCs were studied to demonstrate the working capability of our instrument. The first rate coefficients and product distributions

for the reaction of Ar+ with a variety of aromatic hydrocarbons are then reported through the temperature range of 298-423 K. II. Experimental Methods Flowing Afterglow (FA). The FA method has been well described in the literature,20 and only a brief description follows. A general diagram of the entire apparatus is presented in Figure 1.40 One unique feature of our instrument is that the VTFA is complemented with a variable-temperature neutral flow reactor41,42 and a SIFT. Ions in the FA are generated via electron ionization (EI) and traverse the length (∼1 m) of a stainless steel flow tube via a constant flow (typically 100 m/s) of a helium buffer gas. The helium bath gas (Praxair, >99.995%) was purified by passing it through a copper coil, filled with 4 Å molecular sieves, which was immersed in a liquid nitrogen bath. The helium was preequilibrated to the desired temperature by flowing the gas through a 40 ft copper coil submerged in a constant-temperature bath, prior to entering the FA at the EI source. The effective temperature range of the VTFA is 77-533 K. A blunt nose-cone skimmer (0.5 mm orifice) samples a small amount of the gas flow, which passes through two differentially pumped chambers (Edwards 6 in. diffusion pumps) via a set of electrostatic lenses and a 5/8 in. quadrupole (Extrel). Product ions are detected by an electron multiplier (DeTech) in pulsecounting mode. The detection system has unit mass resolution. To determine reaction kinetics, seven (fixed) inlets in the ∼0.5 m reaction zone are available for the addition of neutral reagents to the ion flow. Most of the neutral reagents were purified by triplicate freeze-pump-thaw cycles. Rate coefficient studies were completed at a variety of temperatures and pressures, as noted below. The reaction rates were obtained under pseudo-

11320 J. Phys. Chem. A, Vol. 104, No. 48, 2000

Cohen et al.

Figure 2. Detailed schematic diagram of the SIFT portion of the instrument which depicts the electrostatic quadrupole deflector (QD) which is used to bend the ions 90° into the mass-selection quadrupole.

first-order conditions with the neutral reagent in large excess. Primary product distributions, which are presented in the following tables, were obtained by extrapolating the normalized ion yields to the zero flow rate limit of the neutral reagent (except where noted below). Due to the extrapolation procedure, the product yields should be considered to have an absolute accuracy of (10%.43 Selected Ion Flow Tube (SIFT). A schematic diagram of the SIFT chamber and its FA ion source is shown in Figure 2. A novel design for our FA-SIFT,21 which incorporates a

quadrupole deflector (QD), was chosen in an effort to reduce the effect that the neutral gas load would have on the sensitivity of the mass selection quadrupole. Ions in the FA source are generated by a precursor gas that is introduced via movable inlets into the flowing helium bath gas which then passes over an EI filament. A 400 cfm Roots blower (Stokes) exhausts most of the gas load introduced into the ion source region. A small portion of the gas flow is sampled into the low-pressure SIFT chamber through a blunt molybdenum orifice (3 mm).7 The SIFT chamber is pumped by a 10 in.

Reactions of Ar+ with Volatile Organic Compounds

J. Phys. Chem. A, Vol. 104, No. 48, 2000 11321

TABLE 1: Variable Temperature Rate Coefficients (k) and Branching Ratios (%) for Calibration Reactions Obtained with the FA Techniquea 298 K

323 K

348 K

373 K

398 K

423 K

lit. rate (298 K)a,b

N2+ + O2 f O2+ + N2

k ) 1.18 ( 0.38

k ) 0.77 ( 0.16

k ) 0.88 ( 0.16

k ) 0.74 ( 0.23

k ) 0.63 ( 0.20

k ) 0.59 ( 0.18

k ) 0.50 ( 0.08

N2+ + CO2 f CO2+ + N2

k ) 10.0 ( 5.2

CO2+ + O2 f O2+ + CO2

k ) 1.14 ( 0.43

k ) 1.12 ( 0.31

k ) 0.96 ( 0.22

k ) 1.26 ( 0.44

k ) 1.06 ( 0.30

k ) 0.59 ( 0.13

k ) 0.55 ( 0.08

N2+ + CH4 f CH2+ + H2 + N2 CH3+ + H + N2 CH4+ + N2

k ) 12.3 ( 5.8 9 77 14

k ) 6.5 ( 0.5 14 81 5

k ) 6.1 ( 0.8 12 83 5

k ) 8.7 ( 1.4 17 80 3

k ) 8.6 ( 1.2 16 81 3

k ) 9.7 ( 1.7 18 79 3

k ) 11.4 ( 1.7

N2+ + CH3OH f CH3+ + OH + N2 CH2OH+ + H + N2 CH3OH+ + N2

k ) 14.1 ( 2.9 53 38 9

k ) 15.2 ( 0.3 63 34 3

k ) 16.2 ( 0.6 63 34 3

k ) 17.2 ( 0.2 63 35 2

k ) 17.6 ( 0.1 63 35 2

k ) 17.9 ( 1.6 61 36 3

k ) 14 ( 4

N2+ + CS2 f CS2+ + N2 S+ + CS + N2

k ) 10.2 ( 3.8 45 55

k ) 12 ( 2

N2+ + CH3CN f CH3+ + CN + N2 CH2CN+ + H + N2 CH3CN+ + N2

k ) 16.8 ( 1.3 5 43 52

k ) 21 ( 6



k ) 8.0 ( 1.6

The rate coefficients are in units of 10-10 cm3/s. b Anicich, V. G. J. Phys. Chem. Ref. Data 1994, 22, 1469.

diffusion pump (Edwards) that is backed by a 37 cfm rotary vane mechanical pump (Stokes). A typical pressure in the SIFT vacuum chamber is 1 × 10-6 Torr when the helium carrier gas is flowing in the source flow tube only. Sampled ions are then focused by six electrostatic lenses within the SIFT chamber. An electrostatic quadrupole deflector22,23,26,27 (QD) is introduced into the ion path immediately before the mass-selection quadrupole in order to bend the ions 90° into the SIFT chamber and away from the neutrals in the plasma. Use of two Faraday cups mounted on the exit lenses of the QD suggests that the ions are bent with >90% efficiency. Ions that are bent 90° by the QD are then focused into the ion selection 5/8 in. quadrupole (Extrel), which is differentially pumped by a 6 in. diffusion pump (Edwards) and backed by the same 37 cfm mechanical pump used for the 10 in. diffusion pump. Mass-selected ions are then focused by 3 electrostatic lenses into a venturi injector, which injects ions from the low-pressure mass selection chamber into the high-pressure reaction flow tube. The design of our venturi injector is very similar to that of Van Doren et al.,11 and is composed of inner and outer annuli with a 2 mm central aperture. Typically, about 40% of the helium flow enters through the inner annulus. As a result of back-streaming, the pressure in the mass-selection chamber is ∼1 × 10-5 Torr with ∼ 0.7 Torr in the reaction flow tube. III. Experimental Calibration The absolute rate coefficients reported here are accurate to within 25%. Unless otherwise noted, at least triplicate measurements were taken for all reactions studied; the errors are one standard deviation of the measurements reported to demonstrate the precision of the measurements. In the discussion and tables, we report the average rate coefficient of these multiple measurements, and the reported error represents one standard deviation. A. FA Calibration Reactions. We have investigated the reactions of a few positive ions with a variety of neutral reagents for which rate coefficients have been studied44 at room

temperature, and a comparison of our values with the literature is presented in Table 1. Non-DissociatiVe Charge-Transfer Reactions. The chargetransfer reaction of N2+ + O2 f N2 + O2+ is important in atmospheric chemistry and has been well studied.45-48 The rate coefficient for this reaction at 298 K has been determined many times, and reported values range from (0.50 ( 0.08) to 2.0 × 10-10 cm3/s.44,47 Our value of (1.2 ( 0.4) × 10-10 cm3/s falls within this range but is higher than the accepted value44 of (0.50 ( 0.08) × 10-10 cm3/s. Our measurement of a faster rate is likely due to the presence of electrons in the FA plasma causing some loss of parent N2+ ion signal due to electron recombination. We have also studied the N2+ + O2 f N2 + O2+ reaction in the temperature range of 298-423 K (Table 1), and have determined a temperature dependence of T(-1.7(0.4). Dunkin et al. 45 reported a temperature dependence of T(-0.6) from 300 to 600 K, while McFarland et al.49 reported a temperature dependence of T(-0.8(0.2) for temperatures e 3560 K. A negative temperature dependence is observed in the range of temperatures studied. The trend is similar to the prior experiments mentioned above; however, the magnitude of the dependence differs. This difference is likely due to electron recombination which occurs in the flow tube. The charge-transfer reaction of CO2+ + O2 has also been studied.44,50,51 The rate coefficient at 298 K, 1.14 × 10-10 cm3/ s, is slightly higher than the accepted value (0.55 × 10-10 cm3/ s). A small negative temperature dependence is observed from 298 to 423 K; however, the uncertainty in the temperature dependence is large. Miller et al. have reported a negative temperature dependence over a wider temperature range (90450 K). DissociatiVe Charge-Transfer Reactions. The reactions of N2+ with CS2, CH3OH, CH3CN, and CH4 proceed by dissociative charge transfer (Table 1). In these cases, some fragmentation after charge transfer occurs and leads to the formation of S+ and CS2+ (from carbon disulfide), CHCN+, CH2CN+, and CH3CN+ (from acetonitrile), CH3+, CH2OH+, and CH3OH+ (from methanol), and CH2+ and CH3+ (from methane). In Table 1, the branching ratios are presented as relative yields of the

11322 J. Phys. Chem. A, Vol. 104, No. 48, 2000 particular product ion (corresponding to the reaction channel indicated), normalized for the sum of all of the observed product ions. Our measured rate coefficient for the N2+ reaction with CS2 is in excellent agreement with the literature. Although our measured rate for the reaction of N2+ with CH3CN is slower than the reported value, there is a large uncertainty in the literature value.44 The rate coefficient at 298 K obtained for the reaction of N2+ with CH3OH, (14.1 ( 2.9) × 10-10 cm3/s, is in excellent agreement with previously published results, (14 ( 4) × 10-10 cm3/s.44 The reaction rate coefficient for N2+ with CH3OH was found to have a slightly positive temperature dependence of T(0.7 ( 0.1) in the range of 298 to 423 K. In agreement with previously published data,44 the primary ionic product observed at all temperatures is CH3+, and the minor ionic products are CH2OH+ and CH3OH+. Similarly, our FA measurement of the N2+ + CH4 reaction rate coefficient, (12.3 ( 5.8) × 10-10 cm3/s, is also in good agreement with the accepted value, (11.4 ( 1.7) × 10-10 cm3/s, although the uncertainty in our measurement is rather large. We also found that the reaction of N2+ with methane has a positive temperature dependence of T(1.7(0.5) between the temperatures of 323 and 423 K. The major ionic product observed in the reaction of N2+ with CH4 is CH3+ at all temperatures, which is in agreement with previous studies.44 However, we also observe CH4+ as a minor product (